POMC and TP53 genetic variability and risk of basal cell carcinoma of skin:
Interaction between host and genetic factors
Cosmeri Rizzato
a,*
, Dominique Scherer
a, Peter Rudnai
b, Eugen Gurzau
c, Kvetoslava Koppova
d,
Kari Hemminki
a,e, Federico Canzian
a, Rajiv Kumar
a, Daniele Campa
aa
German Cancer Research Center DKFZ, Heidelberg, Germany
b
National Institute of Environmental Health, Budapest, Hungary
c
Environmental Health Center, Cluj, Romania
dState Health Institute, Banska Bystrica, Slovakia e
Center for Primary Health Care Research, University of Lund, Malmo¨, Sweden
1. Introduction
Basal cell carcinoma BCC of the skin is the most common neoplasm among the Caucasian population of the western world
[1]. The risk for development of BCC is mainly associated with
environmental factors (especially sun exposure) but also with genetic factors[2].
One of the main risk factors for BCC is ultraviolet (UV) radiation,
through its induction of DNA damage[3–5]. Tanning is acquired
pigmentation that results from exposure to UV, melanin synthesis by cutaneous melanocytes and transport into adjacent keratino-cytes[6].
Melanin production is initiated by
a-melanocyte-stimulating
hormone (a-MSH), which is produced by proteolysis from a multicomponent precursor polypeptide, encoded by the
pro-opiomelanocortin (POMC) gene [7]. The induced POMC/a-MSH
binds to MC1R, which further activates the cyclic adenosine monophosphate (cAMP) signalling system, leading to eumelanin production.
It has been shown that p53 is involved in cell-cycle arrest and
apoptosis in response to UV-induced DNA damage[8,9]; moreover,
A R T I C L E I N F O Article history:
Received 21 December 2010 Received in revised form 4 March 2011 Accepted 10 March 2011
Keywords:
Basal cell carcinoma of the skin p53
POMC
Single nucleotide polymorphism
A B S T R A C T
Background: Basal cell carcinoma (BCC) of the skin is the most common neoplasm among the Caucasian population of the western world. Ultraviolet (UV) radiation-induced p53 activation promotes cutaneous pigmentation by increasing transcriptional activity of pro-opiomelanocortin (POMC) in the skin. Induction of POMC/a-melanocyte-stimulating hormone (a-MSH) activates the melanocortin 1 receptor (MC1R), resulting in skin pigmentation. The tumor suppressor p53 is a key player in stress responses that preserve genomic stability, responding to a variety of insults including DNA damage, hypoxia, metabolic stress and oncogene activation. Malfunction of the p53 pathway is an almost universal hallmark of human tumors. Polymorphisms in the gene encoding p53 (TP53) alter its transcriptional activity, which in turn may influence the UV radiation-induced tanning response.
Objective: The aim of the present work is to test association between POMC and TP53 genetic variability, the possible interplay with host factors and the risk of basal cell carcinoma of skin.
Methods: We covered the variability of the two genes we used 17 tagging polymorphisms in 529 BCC cases and 532 healthy controls. We have also tested the possible interactions between the genetic variants and three known risk factors for BCC: skin complexion, sun effect and skin response to sun exposure.
Results: We did not observe any statistically significant association between SNPs in these two genes and BCC risk overall, nor interactions of SNPs with known BCC risk factors. However we found that, in the group of subjects with lower sun exposure, carriers of one copy of the C allele of the TP53 SNP rs12951053 had a decreased risk of BCC (OR = 0.28, 95% CI 0.12–0.62, P = 0.002).
Conclusions: We have observed that the interplay of an environmental risk factor and one polymorphism in TP53 gene could modulate the risk of BCC.
ß2011 Japanese Society for Investigative Dermatology. Published by Elsevier Ireland Ltd. All rights reserved.
* Corresponding author at: Genomic Epidemiology Group, German Cancer Research Center DKFZ, Im Neuenheimer Feld 280, 69120 Heidelberg, Germany. Tel.: +49 6221 421814; fax: +49 6221 421810.
E-mail address:c.rizzato@dkfz.de(C. Rizzato).
Contents lists available atScienceDirect
Journal of Dermatological Science
j o u r n a l h o m e p a g e : w w w . e l s e v i e r . c o m / j d s
0923-1811/$36.00 ß 2011 Japanese Society for Investigative Dermatology. Published by Elsevier Ireland Ltd. All rights reserved.
Cui et al. [10] showed that the tumor-suppressor protein p53 promotes cutaneous pigmentation following UV irradiation by direct transcriptional activation of POMC in the skin and that p53 absence ablates the tanning response.
TP53 mutations have been detected in about half of all BCCs. Furthermore, it was found that aggressive forms of BCC are
significantly associated with increased p53 expression[1,11].
This tumor suppressor gene is highly polymorphic: so far over
200 SNPs have been identified http://p53.free.fr;
http://www-p53.iarc.fr/ [12]. Few of the many TP53 polymorphisms have been assessed for altered biochemical and/or biological function, or for
their effects on cancer risk in population studies [13]. The
association between p53 codon 72 polymorphism and BCC risk
has been investigated but the findings are controversial[14–17].
Another study investigated the association between BCC cancer
risk and one SNP in the POMC gene with negative results[18].
In this study we investigated the effect of tagging and functional polymorphisms in the entire TP53 and POMC genes on well characterized BCC cases and controls. We analyzed TP53 and POMC jointly with MC1R variants, which had been previously genotyped
in the same cases and controls[19].
2. Materials and methods 2.1. Study population
A set of newly diagnosed cases and controls were recruited as part of a large study on risk of various cancers due to environmental arsenic exposure in Hungary, Romania and Slovakia
between 2002 and 2004[20]. The recruitment was carried out in
the counties of Bacs, Csongrad and Jasz-Nagykun-Szolnok in Hungary; Bihor and Arad in Romania and Nitra in Slovakia. Skin cancer cases (n = 529) were invited on the basis of
histopatholog-ical examinations by pathologists. Hospital-based controls
(n = 532) were included in the study, subject to fulfillment of a set of criteria. All general hospitals in the study area were involved in the process of control recruitment and a rotation scheme was
used in order to achieve appropriate geographical distribution. The controls were surgery, orthopedic and trauma patients with conditions such as appendicitis, abdominal hernias, duodenal ulcers, cholelithiasis and fractures; patients with malignant tumors, diabetes and cardiovascular diseases were excluded. The matching of the controls was done using this criteria: being of the same gender of the index case belonging to the same 5 years age band (30–34, 35–39, etc.), being of the same area. Moreover cases and controls were recruited among those individuals that have resided in the study area for at least one year during their lifetime)[21].
Subsequent to the signing of consent forms by the participants, clinicians took venous blood from cases and controls. The blood samples were kept deep frozen at 80 8C until analysis. A general questionnaire was completed by trained personnel after an interview of the recruited cases and controls. The questionnaire was designed to include information on individual cumulative sun exposure in summer, tanning, skin-complexion, effects of sun-exposure on skin and age/s at diagnosis of BCC; the Fitzpatrick classification was not used because of non-availability of facilities uniformly across all recruiting centers of the participating countries. In addition, the interviews included items on demo-graphic, life-style, socio-economic, medical history, occupational exposures, drinking and nutritional habits, as well as detailed residential history. Ethnic background for the cases and controls was recorded along with other characteristics of the study population. Local ethical boards approved the study plan and design.
2.2. Selection of polymorphisms
We aimed at surveying the entire set of common genetic variants in TP53 and POMC genes. To this end we followed a hybrid functional/tagging approach using the algorithm described by
Carlson and co-workers[22]. All polymorphisms in the region of
TP53 gene including 5 kb upstream of the first exon and 5 kb downstream of the last exon with minor allele frequency MAF 5%
[()TD$FIG]
Fig. 1. Linkage disequilibrium (LD) plot across the TP53 locus, r2
in Caucasians from the International HapMap Project version 26 (http://www.hapmap.org) were included. Tagging SNPs were selected with the use of the Tagger program within Haploview
(http://www.broadinstitute.org/mpg/haploview/; http://
www.broadinstitute.org/mpg/tagger/)[23,24], using pairwise tag-ging with a minimum r2of 0.8 (Fig. 1). We also included additional
tagging SNPs selected among polymorphisms detected in a
previous study[12]. We forced in the selection of tagging SNPs
also rs1042522 Pro72Arg, which is a non-synonymous variant possibly implicated in risk of several cancer types. The SNP rs2909430 tags two other polymorphisms already found associat-ed with cancer risk, a 16 bp duplication in the intron 3 rs17878362
[25]and a SNP in intron 6 rs1625895[25].
SNPs selection for POMC has been described in a previous work
[26]in which the selection has been done using an
haplotype-tagging approach according to the method of Stram et al.[27], with criterion r2
H¼ 0:7. 2.3. Genotyping
The order of DNAs from cases and controls was randomized on PCR plates in order to ensure that an equal number of cases and controls could be analyzed simultaneously. All the genotyping was carried out using the Taqman assay. The MGB Taqman probes and primers were synthesized by Applied Biosystems (Foster City, CA, USA). PCRs were performed according to the manufacturer’s instructions. PCR plates were read on an ABI PRISM 7900HT instrument by Applied Biosystems. MC1R gene has been geno-typed by sequencing the amplification and sequencing conditions
have been described previously[19].
2.4. Statistical analysis
The frequency distribution of genotypes was examined for cases and controls. Hardy–Weinberg equilibrium was tested in cases and in controls separately. The host factors, skin complexion and skin response to sun-exposure were categorized into high (H: light complexion or burns/blisters, respectively), medium (M: medium complexion or mild burns) and low (L: dark complexion or tan/no change) risk groups. Sun exposure was estimated by taking a mean of eight categorical variables measuring average daily exposure to the sun in over the respondents’ lifetimes. The calculated mean was, then, divided in four categories correspond-ing to the hours of sun exposure durcorrespond-ing summer. The four cut-off points were <2.5 h/2.5–3.5 h/3.6–4.5 h/>4.5 h. More detailed
information has been previously reported[19].
We used logistic regression for multivariate analyses to assess the main effects of the genetic polymorphism on BCC risk. The primary end point of the analysis was cancer risk, measured with odds ratio and associated confidence intervals. All estimates were adjusted for age at diagnosis, gender, nationality and risk categories. Considering the large number of comparison performed, we calculated for each gene the number of effective independent
variables, Meff, using the SNP Spectral Decomposition approach
[28]. We obtained a gene-wide Meffvalue for each gene and also a
study-wide Meffvalue, by adding up the gene Meff’s. The
study-wide Meff value was 12, therefore we applied a study-wise
threshold of significance of P = 0.0042 (0.05/12), in order to interpret P-values in light of the multiple comparisons.
We analyzed associations of SNPs with BCC risk by grouping cases according to age, skin complexion, sun effect and skin response to sun exposure, we analyzed a model with the main effects for each SNP and the covariate of interest and a model where the SNP was parameterized nested within the covariate categories; we then computed the likelihood ratio test between the two models (heterogeneity test). Age was divided in 4 quartiles
(<55 years, 55–63 years, 64–72, >72). The host factor subgroups have been described above.
Logistic regression analyses and likelihood ratio test were done with STATA software (StataCorp, College Station, TX, USA).
The haplotype frequencies in cases and controls were inferred with the SAS/Genetics software module (SAS Institute Inc., Cary, NC, U.S.A.) using the expectation–maximization algorithm to generate maximum likelihood estimates. Samples missing one genotype or more were removed from haplotype analyses. 2.5. Gene–gene and gene–environment interactions
We analyzed all the possible pair-wise interactions between SNPs (gene–gene; G–G) and between SNPs and two host factors (skin complexion and skin response to sun exposure) and one environmental factor (sun exposure) (gene-environment; G–E). Assessment of G–G and G–E interactions was carried out using Multifactor Dimensionality Reduction (MDR). The details of MDR
are described elsewhere[29–31]. Briefly, MDR is a data reduction
approach that seeks to identify combinations of multilocus genotypes and discrete environmental factors that are associated with either high risk or low risk of disease. MDR defines a single variable that incorporates information from several loci and/or environmental factors. This new variable can be evaluated for its ability to classify and predict outcome risk status using cross-validation and permutation testing. The MDR software is
open-source and freely available fromhttp://www.epistasis.org.
2.6. Bioinformatics analysis
Potential binding sites of transcription factors within the sequence encompassing the significantly associated SNP were
performed with MatInspector Professional http://genomatix.de/
cgi-bin/matinspector_prof/mat_fam.pl [32]. 3. Results
In this study 529 cases with BCC and 532 controls were recruited from Hungary, Romania and Slovakia. The mean age at
Table 1
Distribution of BCC cases and controls for different characteristics.
Variable Cases (%) Controls (%) P-valuea
Male 237 (44.8) 274 (51.4)
Female 292 (55.2) 259 (48.6)
Mean age (standard deviation) 64.8 (10.3) 60.0 (11.8) Median age (25–75% percentile) 67 (58–73) 61 (52–70) Nationality Hungarian 208 (39) 283 (53) Romanian 125 (24) 118 (22) Slovak 184 (35) 121 (23) Others 12 (2) 10 (2) Skin complexion <0.0001 Light 280 (53) 212 (40) Medium 233 (44) 261 (49) Dark 16 (3) 59 (11)
Skin response to sun-exposure 0.04
Blistered/burnt 185 (35) 141 (26)
Mild burn 169 (32) 159 (30)
Tanning/no change 175 (33) 232 (44)
Average cumulative sun exposure (h per day during summer)b
0.59 <2.4 129 (24) 137 (26) 2.5–3.5 151 (29) 153 (29) 3.6–4.5 135 (26) 111 (21) >4.5 112 (21) 125 (23) a
P-value is for the effect of factor on BCC risk.
b
Sun exposure estimated by calculating a mean of eight categorical variables measuring average daily exposure to the sun over the respondents’ lifetimes. For two cases and six controls exposure information was not available.
diagnosis of the cases (237 men and 292 women) was 64.8 (10.3) years (median 67; 25–75% percentile 58–73) and that of controls (274 men and 259 women) was 60.0 (11.8) years (median 61; 25–75% percentile 52–70). While the complexion and nature of skin response to sun exposure showed association with BCC risk, the average cumulative sun-exposure was not associated with the risk Seventeen SNPs in TP53 and POMC genes were genotyped in each subject of the study and analyzed jointly with genotypes of MC1R SNPs we generated in a previous study[15]. Baseline characteristics of cases and controls are reported inTable 1. Genotype success rate for cases and controls was greater than 95%. Blinded duplicate samples (6.2%) included for quality control showed genotype concordance >99%. The
genotype frequencies for all SNPs were in accordance with Hardy– Weinberg equilibrium in controls, and any deviation from the expected was not statistically significant (data not shown).The distribution of the genotypes and their odds ratios (ORs) for
association with BCC risk are shown inTables 2 and 3. The genotype
frequencies of all SNPs were not found to be significantly different between cases and controls.
3.1. G–G, G–E interactions and subgroup analysis
We have thoroughly analyzed all the possible pair-wise G–G and G–E interactions between the selected SNPs in TP53, POMC and
Table 2
Associations between tagging SNPs in the TP53 gene region and BCC risk. Genotypes Positiona Cases (n = 529)b Controls (n = 532)b ORc 95% CIc p-Value P trend rs1642763 7,498,144 ATP1B2, exon 4 0.28 GG 325 343 1 (ref) AG 169 145 1.15 (0.87–1.53) 0.322 AA 24 25 1.06 (0.58–1.96) 0.845 AG + AA 1.14 (0.87–1.50) 0.338 rs1050540 7,501,467 ATP1B2, 30UTR CC 212 224 1 (ref) 0.56 CT 239 224 1.09 (0.83–1.44) 0.542 TT 67 67 0.96 (0.64–1.45) 0.859 CT + TT 1.06 (0.82–1.37) 0.662 rs1050541 7,501,560 ATP1B2, 30UTR GG 137 143 1 (ref) 0.93 GT 251 239 1.16 (0.85–1.59) 0.338 TT 125 133 1.05 (0.73–1.49) 0.802 GT + TT 1.12 (0.84–1.50) 0.438 rs1641510 7,502,221 intergenic AA 184 171 1 (ref) 0.46 AG 236 247 0.93 (0.70–1.24) 0.628 GG 100 104 0.94 (0.66–1.35) 0.742 AG + GG 0.93 (0.71–1.22) 0.62 rs8073498 7,510,423 intergenic AA 184 198 1 (ref) 0.19 AC 247 248 1.06 (0.80–1.40) 0.682 CC 88 73 1.20 (0.81–1.76) 0.363 AC + CC 1.09 (0.84–1.43) 0.512 rs12951053 7,518,132 TP53, intron 7 AA 440 437 1 (ref) 0.55 AC 79 79 0.97 (0.68–1.39) 0.872 CC 3 7 0.38 (0.09–1.65) 0.198 AC + CC 0.92 (0.65–1.31) 0.656 rs2909430 7,519,370 TP53, intron 4 TT 392 387 1 (ref) 0.96 TC 103 112 1.01 (0.73–1.39) 0.950 CC 15 10 1.52 (0.64–3.60) 0.340 TC + CC 1.05 (0.77–1.43) 0.741 rs9895829 7,519,404 TP53, intron 4 GG 390 388 1 (ref) 0.99 GA 106 117 0.99 (0.72–1.36) 0.960 AA 13 8 1.65 (0.65–4.18) 0.292 GA + AA 1.04 (0.76–1.40) 0.819 rs1042522 7,520,197 TP53, exon 3 CC 292 297 1 (ref) 0.91 CG 186 178 1.09 (0.83–1.44) 0.523 GG 40 46 0.91 (0.57–1.46) 0.702 CG + GG 1.06 (0.82–1.37) 0.679 rs12602273 7,523,738 TP53, intron 1 CC 448 441 1 (ref) 0.52 CG 68 78 0.85 (0.59–1.23) 0.385 GG 3 2 1.31 (0.19–9.2) 0.789 CG + GG 0.86 (0.6–1.24) 0.418 rs2287499 7,532,893 WRAP53, exon 1 CC 406 397 1 (ref) 0.19 CG 107 114 0.93 (0.68–1.27) 0.631 GG 5 12 0.46 (0.15–1.39) 0.169 CG + GG 0.88 (0.65–1.2) 0.43 a
Position of SNP on chromosome 17, in base pairs referred to UCSC Genome Browser on Human Mar. 2006 Assembly. In parentheses we report the position of the polymorphism with respect to the gene. Some SNPs are located in genes immediately flanking TP53 at the 30and 50.
b
Numbers may not add up to 100% of subjects due to genotyping failure.
c
MC1R by use of the MDR approach. Analysis using the MDR method with 10-fold cross validation did not reveal any statistically significant interaction (the best model described had cross validation consistency 9/10, test balance accuracy 0.56, OR 1.61 CI 0.74–3.49, P value = 0.23). Analyzing the data in subgroups stratifying for age, gender and the host/environmental factors (skin complexion, sun exposure and skin response to sun exposure), we found that two polymorphic variants of TP53 had a study-wise (P-value <0.0041) significant association with BCC risk. Results of subgroup analysis for rs12951053 and rs8073498 are reported in Table 4. Carriers of one copy of the C allele of rs12951053 had a decreased risk of BCC in the group of subjects with lower sun exposure during the summer (OR = 0.28, 95% CI 0.12–0.62, P = 0.002). Moreover, carriers of the C allele of the same polymorphism show an increased risk in a group of subject with age ranged between 55 and 63 (OR = 2.67 95% CI 1.27–5.64, P = 0.01) and subjects more prone to sun burns (OR = 1.95, 95% CI 0.99–3.81, P = 0.052), even if these associations did not reach the threshold for multiple comparison analysis. When stratifying for sun exposure, the SNP rs12951053 had a heterogeneity test P value of 0.0187.
We observed an increased risk of BCC in carriers of one copy of the C allele of SNP rs8073498 in TP53 in the group of subjects with intermediate-high sun exposure during the summer (OR = 2.83, 95% CI 1.50–5.35, P = 0.001), although the heterogeneity test was not significant (pheterogeneity= 0.43). Carriers of at least one copy of
C allele show an association in this subgroup (OR = 2.32, 95% CI 1.29–4.16, P = 0.005), even if this association did not reach the threshold for multiple comparison analysis. All the other poly-morphisms, considering the various subgroups, did not show any
statistically significant association. Supplementary Tables 1–4
show the analysis of each polymorphism of this study considering the various subgroups of age and sun-related exposures.
Haplotype analysis did not detect associations with BCC risk
(showed in detail inSupplementary Table 5).
Searches for potential binding sites of transcription factors were performed with MatInspector within 100 bp sequence surround-ing rs12951053 and rs8073498. For the rs12951053 the predicted binding of transcription factor did not differ between the two alleles.
Twenty-one binding sites for transcription factors were detected in the 100 bp sequence surrounding the A allele of rs8073498. Four of those binding sites were abolished in the presence of the C allele (Fig. 2). Two of the abolished sites were for p53 (matrix similarity 0.94 and 0.98) and 2 for Iroquois homeobox transcription factors (not expressed in the skin).
4. Discussion
The risk for development of BCC is mainly associated with environmental factors (especially sun exposure) but also with
genetic factors [2]. In this study we investigated the genetic
variability of TP53 and POMC genes using a tagging approach and selecting 17 SNPs. Using this method we covered all the known common genetic variation of these genes, including
polymorph-isms coming from a recent study[12], but we did not find any
significant difference of genotype and haplotypes distribution between cases and controls. Polymorphisms in TP53 have been
found to be associated with cancer risk in a variety of tissues[33–
35].
Two studies found no association of the p53 codon 72
polymorphism with BCC risk [14,15], while in two others
[16,17]the polymorphism has been found associated.
In particular the same population, in which Han et al. reported an association between the p53 codon 72 polymorphism with tanning response for BCC risk in a prospective cohort of women,
Table 3
Associations between tagging SNPs in the POMC gene region and BCC risk.
Genotypes Positiona Casesb Controlsb OR 95% CI P-value P trend rs13002622 25,223,520 Intron, EFR3B TT 248 236 1 (ref) 0.42 TC 223 237 0.88 (0.67–1.15) 0.363 CC 45 48 0.95 (0.6–1.52) 0.846 TC + CC 0.89 (0.69–1.16) 0.395 rs1866146 25,234,077 30UTR, EFR3B AA 186 194 1 (ref) 0.55 AG 261 269 1.09 (0.82–1.44) 0.551 GG 70 63 1.17 (0.77–1.77) 0.472 AG + GG 1.10 (0.85–1.44) 0.469 rs1042571 25,237,391 30UTR, POMC GG 361 366 1 (ref) 0.56 GA 146 149 0.87 (0.65–1.15) 0.327 AA 14 8 1.96 (0.78–4.97) 0.154 GA + AA 0.92 (0.69–1.21) 0.546 rs7566506 25,272,977 Intergenic CC 447 447 1 (ref) 0.47 CA 67 73 0.97 (0.67–1.41) 0.871 AA 3 5 0.64 (0.14–2.85) 0.560 CA + AA 0.95 (0.66–1.37) 0.775 rs10202360 25,274,369 Intergenic AA 334 327 1 (ref) 0.59 AC 172 175 1.05 (0.80–1.39) 0.711 CC 14 17 0.80 (0.38–1.72) 0.574 AC + CC 1.03 (0.79–1.35) 0.827 rs28932474 25,383,068 Intergenic GG 474 475 1 (ref) 0.58 GC 48 54 0.86 (0.56–1.32) 0.479 CC 0 0 a
Position of SNP on chromosome 2, in base pairs referred to UCSC Genome Browser on Human Mar. 2006 Assembly. In parentheses we report the position of the polymorphism with respect to the gene.
b
Table 4
Association of rs12951053 and rs8073498 with BCC risk by subgroups of environmental and host factors.
Cases Controls Per allele P value AA vs. CC P value AA vs. (AC + CC) P value P trend
AA AC CC AA AC CC ORa 95% CIa ORa 95% CIa ORa 95% CIa rs12951053 Sun-exposure (Pheterogeneity= 0.0187)b >2.4 120 11 0 108 30 0 0.28 (0.12–0.62) 0.002 2.4–3.5 96 17 2 104 9 2 1.91 (0.77–4.77) 0.16 0.62 (0.07–5.52) 0.67 1.66 (0.71–3.89) 0.25 0.19 3.6–4.5 105 19 1 83 14 3 1.11 (0.51–2.42) 0.78 0.33 (0.03–3.42) 0.35 0.99 (0.47–2.06) 0.97 0.58 4.5 119 32 0 142 26 2 1.36 (0.75–2.48) 0.31 1.28 (0.71–2.31) 0.41 0.44
Skin response to sun effectc
Blistered/burnt 147 36 1 122 15 1 1.95 (0.99–3.81) 0.05 0.76 (0.05–12.69) 0.85 1.87 (0.97–3.59) 0.06 0.06 Mild burn 144 21 2 125 27 5 0.69 (0.36–1.33) 0.27 0.37 (0.06–2.30) 0.29 0.65 (0.35–1.20) 0.17 0.08 Tanning/no change 138 22 0 183 35 1 0.83 (0.46–1.52) 0.55 0.81 (0.44–1.46) 0.48 0.41 Riskc Low 140 22 0 189 37 2 0.81 (0.45–1.46) 0.48 0.76 (0.42–1.38) 0.37 0.26 Medium 63 12 1 67 17 1 0.81 (0.35–1.90) 0.63 1.57 (0.08–29.54) 0.76 0.85 (0.37–1.93) 0.69 0.57 High 237 45 2 181 25 4 1.22 (0.71–2.11) 0.47 0.31 (0.05–1.92) 0.21 1.10 (0.65–1.86) 0.72 0.68 Aged <55 82 6 1 139 23 4 0.43 (0.16–1.15) 0.09 0.35 (0.04–3.43) 0.37 0.42 (0.17–1.04) 0.06 0.07 55–63 87 25 0 113 13 1 2.67 (1.27–5.64) 0.01 2.51 (1.2–5.22) 0.01 0.043 64–72 135 26 0 102 21 1 0.93 (0.49–1.77) 0.83 0.88 (0.47–1.66) 0.69 0.60 >72 136 22 2 83 22 1 0.61 (0.31–1.22) 0.16 1.10 (0.08–15.23) 0.95 0.63 (0.33–1.24) 0.18 0.22 Complexionc Light 235 40 2 174 31 4 0.87 (0.51–1.48) 0.60 0.27 (0.04–1.73) 0.17 0.80 (0.48–1.34) 0.39 0.46 Medium 191 36 1 211 41 2 0.96 (0.58–1.60) 0.88 0.73 (0.06–8.73) 0.80 0.95 (0.58–1.57) 0.85 0.77 Dark 13 3 0 51 7 1 1.90 (0.41–8.75) 0.41 1.62 (0.36–7.27) 0.53 0.76 p53rs8073498 Sun-exposureb (Pheterogeneity= 0.4322) >2.4 51 59 20 44 78 17 0.59 (0.33–1.05) 0.07 0.77 (0.34–1.76) 0.540 0.63 (0.36–1.08) 0.09 0.59 2.4–3.5 39 58 18 39 58 16 1.00 (0.54–1.84) 0.10 0.95 (0.40–2.25) 0.911 0.99 (0.55–1.77) 0.97 0.82 3.6–4.5 38 69 18 48 34 17 2.83 (1.50–5.35) 0.001 1.39 (0.61–3.18) 0.431 2.32 (1.29–4.16) 0.005 0.10 4.5 56 61 32 67 78 23 0.93 (0.56–1.55) 0.79 1.57 (0.80–3.08) 0.185 1.08 (0.67–1.73) 0.75 0.21
Skin response to sun effectc
Blistered/burnt 63 88 30 52 65 21 1.05 (0.63–1.74) 0.86 1.09 (0.54–2.18) 0.816 1.06 (0.65–1.71) 0.82 0.59 Mild burn 61 86 21 50 80 24 0.89 (0.54–1.49) 0.67 0.69 (0.33–1.44) 0.323 0.85 (0.52–1.37) 0.50 0.35 Tanning/no change 57 68 34 92 98 28 1.24 (0.77–2.00) 0.37 1.68 (0.90–3.13) 0.105 1.35 (0.87–2.1) 0.18 0.05 Riskc Low 57 70 34 94 105 28 1.19 (0.75–1.90) 0.47 1.72 (0.92–3.20) 0.088 1.31 (0.85–2.03) 0.22 0.04 Medium 30 38 9 29 41 13 1.01 (0.50–2.03) 0.99 0.58 (0.20–1.65) 0.309 0.89 (0.46–1.74) 0.73 0.45 High 97 139 45 75 102 32 0.98 (0.65–1.48) 0.91 1.07 (0.60–1.89) 0.824 1.00 (0.67–1.48) 0.99 0.74 Aged <55 28 44 15 67 76 23 1.56 (0.85–2.86) 0.15 1.34 (0.59–3.04) 0.488 1.50 (0.85–2.65) 0.17 0.21 55–63 40 54 18 49 64 12 0.97 (0.55–1.72) 0.92 1.77 (0.74–4.23) 0.197 1.10 (0.64–1.88) 0.74 0.25 64–72 58 80 24 40 64 19 0.83 (0.49–1.42) 0.50 0.82 (0.39–1.73) 0.607 0.83 (0.5–1.38) 0.47 0.63 >72 58 69 31 42 44 19 1.13 (0.64–2.01) 0.68 1.26 (0.61–2.59) 0.539 1.17 (0.69–1.98) 0.57 0.60 Complexionb Light 105 129 42 79 98 30 1.04 (0.69–1.58) 0.84 0.99 (0.56–1.77) 0.979 1.03 (0.70–1.52) 0.88 0.89 Medium 74 111 41 92 124 36 1.10 (0.73–1.66) 0.65 1.28 (0.73–2.25) 0.390 1.14 (0.77–1.69) 0.51 0.23 Dark 5 7 4 26 26 7 1.46 (0.40–5.37) 0.57 2.58 (0.51–12.92) 0.250 1.74 (0.52–5.75) 0.37 0.19
aOR: odds ratio; CI: confidence interval. Study-wise significant associations (P < 0.0042) are written in bold, associations with P < 0.05 but not reaching the threshold for
multiple comparison are underlined.
b
Adjusted for age, gender, risk (which consists of the combination of skin complexion and skin response to sun exposure) and nationality.
c
Adjusted for age, gender and nationality.
d
Adjusted for gender, risk and nationality.
[()TD$FIG]
Fig. 2. Output of MatInspector software. The yellow arrow indicates the position of rs8073498 SNP. The C allele abolishes recognition motifs for p53 (in a green frame). The recognition motif is predicted with a matrix similarity of 0.94 and 0.98 (a. output for A allele, b. output for C allele). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of the article.)
showed no association between one SNP in POMC gene
(rs1042571) and BCC risk[18]. Our results show no association
with the p53 codon 72 polymorphism, and confirm lack of association with POMC polymorphisms covering the whole gene,
including rs1042571. Nan et al. [36]tested also an interaction
analysis between p53 codon 72 polymorphism and MC1R variants. The authors found that there was an association between the two genes and melanoma risk but not for BCC risk. Our results on the interactions between the two gene variants confirm in a bigger independent population their null finding.
Sun exposure is the major etiological factor in the genesis of BCC; however, many studies have suggested that risk may involve an interplay between genetic (SNPs), host (skin complexion, skin response to sun exposure) and environmental (sun exposure) factors
[2]. Thus we have thoroughly analyzed all the possible interactions
and associations between the selected polymorphisms and known risk factors for BCC. Analyzing the data in subgroups stratifying for age, gender and the host factors previously mentioned, we found that two polymorphisms were associated with BCC risk. We found that, in the group of subjects with lower sun exposure during the summer, carriers of one copy of the C allele of the rs12951053 had a decreased risk of BCC and subjects with intermediate-high sun exposure had an increased risk of BCC in carriers of one copy of the C allele of SNP rs8073498, although the heterogeneity test in this subgroup analysis was not significant. The rs12951053 polymor-phism has been found associated with increased risk in other cancer
types[37–39]. On the other hand, we observed it to be associated
with a decrease in BCC risk. p53 is known to act in many different ways, depending on the tissue and environmental stimuli; in this specific case its function is likely related with UV response. In different cancer types the major function can be a different one with different regulations. In silico analysis predicts that several variants in moderate LD (0.5 < r2<0.7) with TP53 rs12951053 can affect a
transcription factor binding site; in addition, rs2287498, which is in
the flanking gene WDR79 and in LD with rs12951053 (r2= 0.62), is
predicted to affect function at a splice site[39]. The mRNA encoding this protein plays a critical role in the regulation of p53 expression at the post-transcriptional level; it is involved both in maintaining
basal p53 mRNA levels and in p53 induction upon DNA damage[40].
So far there is no published association between genotypes in rs8073498 and cancer risk. We assessed the putative effect of the C allele by the MatInspector program, which predicts that this variant may abolish recognition motifs for p53 itself (Fig. 2).
In this study we had more than 80% power to detect a possible association with a minimum OR of 1.28 for a SNP with minor allele frequency of 0.45 in the controls assuming alpha = 0.05, two-sided test and a codominant model. In conclusion, we have observed that the interplay of sun exposure and one polymorphism in TP53 gene association could modulate the risk of BCC. However these results have to be taken with caution due to the relatively large number of comparisons done and have to be replicated in a larger independent study.
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in the online version, atdoi:10.1016/j.jdermsci.2011.03.006.
References
[1] Miller DL, Weinstock MA. Nonmelanoma skin cancer in the United States: incidence. J Am Acad Dermatol 1994;30(5, Part 1):774–8.
[2] Kumar R. Molecular epidemiology of skin cancer. In: Ringborg U, Brandeber Y, Breitbart E, et al., editors. Prevention of skin cancer. London: Informa Health-care; 2006. p. 225–40.
[3] Armstrong BK, Kricker A, English DR. Sun exposure and skin cancer. Australias J Dermatol 1997;38(Suppl. 1):S1–6.
[4] Brash DE. Sunlight and the onset of skin cancer. Trends Genet 1997;13:410–4. [5] English DR, Armstrong BK, Kricker A, Fleming C. Sunlight and cancer. Cancer
Causes Control 1997;8:271–83.
[6] Lin JY, Fisher DE. Melanocyte biology and skin pigmentation. Nature 2007;445:843–50.
[7] Schauer E, Trautinger F, KA˜ ôck A, Schwarz A, Bhardwaj R, Simon M, et al. Proopiomelanocortin-derived peptides are synthesized and released by hu-man keratinocytes. J Clin Invest 1994;93:2258–62.
[8] Hussain SP, Hollstein MH, Harris CC. p53 tumor suppressor gene: at the crossroads of molecular carcinogenesis molecular epidemiology, and human risk assessment. Ann N Y Acad Sci 2000;919:79–85.
[9] Olivier M, Eeles R, Hollstein M, Khan MA, Harris CC, Hainaut P. The IARC TP53 database: new online mutation analysis and recommendations to users. Hum Mutat 2002;19:607–14.
[10] Cui R, Widlund HR, Feige E, Lin JY, Wilensky DL, Igras VE, et al. Central role of p53 in the suntan response and pathologic hyperpigmentation. Cell 2007;128:853–64.
[11] Tilli CM, Van Steensel MA, Krekels GA, Neumann HA, Ramaekers FC. Molecular aetiology and pathogenesis of basal cell carcinoma. Br J Dermatol 2005;152:1108–24.
[12] Garritano S, Gemignani F, Voegele C, Nguyen-Dumont T, Le Calvez-Kelm F, De Silva D, et al. Determining the effectiveness of high resolution melting analysis for SNP genotyping and mutation scanning at the TP53 locus. BMC Genet 2009;10:5.
[13] Whibley C, Pharoah PD, Hollstein M. p53 polymorphisms: cancer implications. Nat Rev Cancer 2009;9:95–107.
[14] Almquist LM, Karagas MR, Christensen BC, Welsh MM, Perry AE, Storm C, et al. The role of TP53 and MDM2 polymorphisms in TP53 mutagenesis and risk of non-melanoma skin cancer. Carcinogenesis 2011;32:327–30.
[15] Bendesky A, Rosales A, Salazar AM, Sordo M, Peniche J, Ostrosky-Wegman P. p53 codon 72 polymorphism, DNA damage and repair, and risk of non-melanoma skin cancer. Mutat Res 2007;619:38–44.
[16] Chen YC, Xu L, Guo YL, Su HJ, Hsueh YM, Smith TJ, et al. Genetic polymorphism in p53 codon 72 and skin cancer in southwestern Taiwan. J Environ Sci Health A Tox Hazard Subst Environ Eng 2003;38:201–11.
[17] Han J, Cox DG, Colditz GA, Hunter DJ. The p53 codon 72 polymorphism, sunburns, and risk of skin cancer in US Caucasian women. Mol Carcinogen 2006;45:694–700.
[18] Nan H, Kraft P, Hunter DJ, Han J. Genetic variants in pigmentation genes, pigmentary phenotypes, and risk of skin cancer in Caucasians. Int J Cancer 2009;125:909–17.
[19] Scherer D, Bermejo JL, Rudnai P, Gurzau E, Koppova K, Hemminki K, et al. MC1R variants associated susceptibility to basal cell carcinoma of skin: interaction with host factors and XRCC3 polymorphism. Int J Cancer 2008;122:1787–93. [20] Lindberg AL, Goessler W, Gurzau E, Koppova K, Rudnai P, Kumar R, et al. Arsenic exposure in Hungary, Romania and Slovakia. J Environ Monit 2006;8:203–8.
[21] Thirumaran RK, Bermejo JL, Rudnai P, Gurzau E, Koppova K, Goessler W, et al. Single nucleotide polymorphisms in DNA repair genes and basal cell carcino-ma of skin. Carcinogenesis 2006;27:1676–81.
[22] Carlson CS, Eberle MA, Rieder MJ, Yi Q, Kruglyak L, Nickerson DA. Selecting a maximally informative set of single-nucleotide polymorphisms for association analyses using linkage disequilibrium. Am J Hum Genet 2004;74:106–20. [23] Barrett JC, Fry B, Maller J, Daly MJ, Haploview:. analysis and visualization of LD
and haplotype maps. Bioinformatics 2005;21:263–5.
[24] de Bakker PI, Yelensky R, Pe’er I, Gabriel SB, Daly MJ, Altshuler D. Efficiency and power in genetic association studies. Nat Genet 2005;37:1217–23. [25] Lazar V, Hazard F, Bertin F, Janin N, Bellet D, Bressac B. Simple sequence repeat
polymorphism within the p53 gene. Oncogene 1993;8:1703–5.
[26] Canzian F, Cox DG, Setiawan VW, Stram DO, Ziegler RG, Dossus L, et al. Comprehensive analysis of common genetic variation in 61 genes related to steroid hormone and insulin-like growth factor-I metabolism and breast cancer risk in the NCI breast and prostate cancer cohort consortium. Hum Mol Genet 2010;19:3873–84.
[27] Stram DO, Haiman CA, Hirschhorn JN, Altshuler D, Kolonel LN, Henderson BE, et al. Choosing haplotype-tagging SNPS based on unphased genotype data using a preliminary sample of unrelated subjects with an example from the multiethnic cohort study. Hum Hered 2003;55:27–36.
[28] Gao X, Starmer J, Martin ER. A multiple testing correction method for genetic association studies using correlated single nucleotide polymorphisms. Genet Epidemiol 2008;32:361–9.
[29] Marylyn DR, Lance WH, Jason HM. Power of multifactor dimensionality reduction for detecting gene–gene interactions in the presence of genotyping error, missing data, phenocopy, and genetic heterogeneity. Genet Epidemiol 2003;24:150–7.
[30] Moore JH. Computational analysis of gene–gene interactions using multifactor dimensionality reduction. Expert Rev Mol Diagn 2004;4:795–803. [31] Ritchie MD, Hahn LW, Roodi N, Bailey LR, Dupont WD, Parl FF, et al.
Multifac-tor-dimensionality reduction reveals high-order interactions among estrogen-metabolism genes in sporadic breast cancer. Am J Hum Genet 2001;69: 138–47.
[32] Cartharius K, Frech K, Grote K, Klocke B, Haltmeier M, Klingenhoff A, et al. MatInspector and beyond: promoter analysis based on transcription factor binding sites. Bioinformatics 2005;21:2933–42.
[33] Boldrini L, Gisfredi S, Ursino S, Lucchi M, Greco G, Mussi A, et al. Effect of the p53 codon 72 and intron 3 polymorphisms on non-small cell lung cancer (NSCLC) prognosis. Cancer Invest 2008;26:168–72.
[34] Gemignani F, Moreno V, Landi S, Moullan N, Chabrier A, Gutierrez-Enriquez S, et al. A TP53 polymorphism is associated with increased risk of colorectal cancer and with reduced levels of TP53 mRNA. Oncogene 2004;23:1954–6. [35] Wang-Gohrke S, Weikel W, Risch H, Vesprini D, Abrahamson J, Lerman C, et al.
Intron variants of the p53 gene are associated with increased risk for ovarian cancer but not in carriers of BRCA1 or BRCA2 germline mutations. Br J Cancer 1999;81:179–83.
[36] Nan H, Qureshi AA, Hunter DJ, Han J. Interaction between p53 codon 72 polymorphism and melanocortin 1 receptor variants on suntan response and cutaneous melanoma risk. Br J Dermatol 2008;159:314–21.
[37] Garcia-Closas M, Kristensen V, Langerod A, Qi Y, Yeager M, Burdett L, et al. Common genetic variation in TP53 and its flanking genes WDR79 and ATP1B2, and susceptibility to breast cancer. Int J Cancer 2007;121:2532–8. [38] Koshiol J, Hildesheim A, Gonzalez P, Bratti MC, Porras C, Schiffman M, et al.
Common genetic variation in TP53 and risk of human papillomavirus persis-tence and progression to CIN3/cancer revisited. Cancer Epidemiol Biomarkers Prev 2009;18:1631–7.
[39] Schildkraut JM, Iversen ES, Wilson MA, Clyde MA, Moorman PG, Palmieri RT, et al. Association between DNA damage response and repair genes and risk of invasive serous ovarian cancer. PLoS One 2010;5:e10061.
[40] Mahmoudi S, Henriksson S, Corcoran M, Mendez-Vidal C, Wiman KG, Farnebo M. Wrap53, a natural p53 antisense transcript required for p53 induction upon DNA damage. Mol Cell 2009;33:462–71.
